ppy/OntologySummit2014-s04_chat-transcript_edited_20140206b.txt ------ Chat transcript from room: summit_20140206 2014-02-06 GMT-08:00 [PST] ------ [8:56] PeterYim: Welcome to the = OntologySummit2014 session-04 Track-C: Overcoming Ontology Engineering Bottlenecks-I - Thu 2014-02-06 = Summit Theme: OntologySummit2014: "Big Data and Semantic Web Meet Applied Ontology" Track-C Focus: Overcoming Ontology Engineering Bottlenecks Track-C Co-champions: Professor KrzysztofJanowicz (University of California, San Diego), Professor PascalHitzler (Wright State University), Dr. MatthewWest (Information Junction) Session Topic: Overcoming Ontology Engineering Bottlenecks -- Strategies and Building Blocks Session Co-chairs: Professor KrzysztofJanowicz and Professor PascalHitzler Briefings: * Professor WernerKuhn (University of California, Santa Barbara) - "Abstracting behavior in ontology engineering" * Professor AldoGangemi (University Paris 13) - "Knowledge Patterns as one means to overcome ontology design bottlenecks" * Mr. KarlHammar (Jonkoping University) - "Reasoning Performance Indicators for Ontology Design Patterns" Logistics: * Refer to details on session page at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2014_02_06 * (if you haven't already done so) please click on "settings" (top center) and morph from "anonymous" to your RealName; also please enable "Show timestamps" while there. * Mute control (phone keypad): *7 to un-mute ... *6 to mute * Attn: Skype users ... see: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2014_02_06#nid44NP ** you may connect to (the skypeID) "joinconference" whether or not it indicates that it is online (i.e. even if it says it is "offline," you should still be able to connect to it.) ** if you are using skype and the connection to "joinconference" is not holding up, try using (your favorite POTS or VoIP line, etc.) either your phone, skype-out or google-voice and call the US dial-in number: +1 (206) 402-0100 ... when prompted enter Conference ID: 141184# ** Can't find Skype Dial pad? *** for Windows Skype users: Can't find Skype Dial pad? ... it's under the "Call" dropdown menu as "Show Dial pad" *** for Linux Skype users: if the dialpad button is not shown in the call window you need to press the "d" hotkey to enable it * when posting in this Chat-room, kindly observe the following ... ** whenever a name is used, please use the full WikiWord name format (every time you don't, some volunteer will have to make an edit afterwards) ** always provide context (like: "[ref. JaneDoe's slide#12], I think the point about context is great" ... rather than "that's great!" as the latter would mean very little in the archives.) ** when responding to a specific individual's earlier remarks, please cite his/her full WikiWord names *and* the timestamp (in PST) of his/her post that you are responding to (e.g. "@JaneDoe [11:09] - I agree, but, ...") ** use fully qualified url's (include http:// ) without symbols (like punctuations or parentheses, etc.) right before of after that URL Attendees: AdilaKrisnadhi, AldoGangemi, AliHashemi, AmandaVizedom, AnatolyLevenchuk, AndreaWesterinen, BartGajderowicz, BoYan, BobMorris, BobbinTeegarden, CarmenChui, CarolBean, CharlesVardeman, ChristineKapp, ChristophLange, ConradBeaulieu, DaliaVaranka, DennisWisnosky, DonnaFritzsche, EdBernot, FranLightsom, FrankLoebe, GaryBergCross, HaroldBoley, HelenCouclelis, HensonGraves, JacoDuPreez, JamesOverton, JamesWilson, JensOrtmann, JimSolderitsch, JoelBender, JohnAYanosyJr, KarlHammar, KenBaclawski, KrzysztofJanowicz, LamarHenderson, LeoObrst, LesMorgan, LianaKiff, MarcelaS, MarcelaVegetti, MarkFox, MatthewWest, MeganKatsumi, MichaelGruninger, MichaelRiben, MikeBennett, MikeDean, NancyWiegand, OliverKutz, PascalHitzler, PeterYim, SimonCox, SimonSpero, SundayOjo, TaraAthan, TerryLongstreth, TimFinin, ToddSchneider, TorstenHahmann, UriShani, VictorAgroskin, WernerKuhn, amin, == Proceedings == [7:05] anonymous morphed into ConradBeaulieu [8:39] anonymous morphed into DaliaVaranka [8:49] anonymous morphed into BobMorris [8:57] JacoDP morphed into JacoDuPreez - [9:10] BobMorris: Peter: I running Skype on Ubuntu Linux 12.10; when I log on to joinconference I get an audio prompt to enter the conference ID, but no indication of how/where to do so. Entering it in the skype chat window has no effect. Suggestion? - [9:13] PeterYim: @BobMorris, please check if this helps - http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2014_02_06#nid44NW - [9:15] BobMorris: Sorry; found it between the time of my q and your a. Joined OK. - [9:16] BobMorris: (It needs less tacky music on hold.... :-) ) - [9:17] PeterYim: glad it worked ... (blame the service provider for the music! :) [9:18] anonymous morphed into WernerKuhn [9:21] anonymous morphed into AldoGangemi [9:23] anonymous morphed into KarlHammar [9:23] KrzysztofJanowicz: Hi all, [9:23] KrzysztofJanowicz: we are waiting for peter. [9:23] KrzysztofJanowicz: keep in mind to press *7 before speaking to unmute [9:24] EdBernot: Thanks, will be standing by muted. [9:25] KrzysztofJanowicz: please also keep in mind to mention the slide number every time you change the slide as this is the only way peter (using the VNC) can advance the slides and researchers listening to your talk in the future will be able to sync your slides and audio. - [9:24] KarlHammar: Hi. Is the VNC session supposed to be live yet? - [9:24] KrzysztofJanowicz: VNC should be up very soon - [9:26] KrzysztofJanowicz: VNC is running - [9:28] KarlHammar: I'm having some trouble viewing the VNC. Just getting a blank screen, on both Firefox and Safari. Is it just me? - [9:28] JacoDuPreez: Same here - [9:28] KarlHammar: I'm having some trouble viewing the VNC. Just getting a blank screen, on both Firefox and Safari. Is it just me? - [9:29] PascalHitzler: VNC works on my firefox. - [9:29] PascalHitzler: sorry, no I have it on IE :) - [9:30] KrzysztofJanowicz: I have VNC working on firefox - [9:53] TaraAthan: @Karl - are you behind a firewall? If so, VNC may not work. - [9:54] TaraAthan: The slides may be downloaded directly as PDFs. [9:25] anonymous morphed into MichaelRiben [9:27] anonymous1 morphed into CharlesVardeman [9:28] anonymous morphed into BartGajderowicz [9:30] anonymous morphed into CarmenChui [9:30] anonymous morphed into LesMorgan [9:30] anonymous1 morphed into ConradBeaulieu [9:31] anonymous1 morphed into SimonCox [9:31] anonymous1 morphed into DaliaVaranka [9:33] anonymous morphed into JamesWilson [9:34] PeterYim: == KrzysztofJanowicz starts the session on behalf of the Track-C session co-champions ... see slides under: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2014_02_06#nid45NX [9:35] anonymous morphed into TorstenHahmann [9:35] anonymous1 morphed into BoYan [9:37] anonymous2 morphed into Adila [9:38] Adila morphed into AdilaKrisnadhi [9:39] PeterYim: == WernerKuhn presenting ... [9:40] anonymous2 morphed into LamarHenderson - [9:44] Adila Krisnadhi: What page of Werner's slide are we on right now? - [9:44] PascalHitzler: page 1 :) - [9:44] AdilaKrisnadhi: Thanks! [9:45] anonymous2 morphed into CarolBean [9:47] PeterYim: @anonymous, please morph from "anonymous" to your RealName (via the "Settings" button at top center of this chat window; also please enable "Show timestamps" while there. [9:48] SimonSpero: [Slide 2] : "White Things" as Sortals; is this a valid sortal? [9:49] anonymous2 morphed into HelenCouclelis [9:50] TerryLongstreth: @Simon: I'd say so, if your requirement is to organize things by visibility (albedo) [9:50] anonymous2 morphed into JohnAYanosyJr [9:51] SimonSpero: @TerryLongstreth: when I was reading up on sortals [in the context of Relative Identity] "green thing" was used as the example of non-sortal predicates, as they cannot be counted [9:52] KrzysztofJanowicz: we have to advance the slide - [9:52] SimonCox: 5 [9:54] GaryBergCross: I take the generic aspect - free of rep languages to mean they would be a conceptual model. - [9:56] KrzysztofJanowicz: next slide - [9:57] KrzysztofJanowicz: we are now on slide 7 [9:57] SimonSpero: Slide 6: Duck typing. Better hope it's a duck, or use something non-monotonic [9:58] SimonCox: @Simon - non-monotonic? [9:59] SimonCox: In UML this is <> ? [10:01] SimonSpero: SimonCox: adding more information in a non-monotonic logic may invalidate previous inferences (e.g. birds fly, but penguins don't) [10:02] TaraAthan: Duck typing seems like it fits the reverse example: most flying vertebrates are birds, but some are actually bats. [10:03] LeoObrst: "White thing" (and similar) is sometimes considered a restricted sortal. I myself find that problematic. [10:03] TaraAthan: I thought the point about "composition" was that white wine is not really white. [10:04] PascalHitzler: Duck typing: On a modeling level, I think a key idea here is to use property fillers to identify "what" something is, rather than through direct typing. Makes a lot of sense to me. [10:05] SimonSpero: [Resource] The place to go for haskell info is http://www.haskell.org/haskellwiki/Haskell [10:05] PascalHitzler: And seems to be not at odds with, e.g., using OWL and rules for modeling (at this level, anyway). @Werner: why try introduce a functional syntax, i.e. what's the additional value of going there? - [10:05] KrzysztofJanowicz: Slide 11 [10:06] MatthewWest: The use of behaviours looks a bit like a facet analysis approach. [10:06] SimonSpero: @Pascal: S-Expr notation has some good points cough kif [10:07] GaryBergCross: @PACAL Comment at [13:04 on property fillers for identity sounds Aristotelian on necessary properties. [10:07] TaraAthan: In another language with functional programming support, Scala, "duck typing" is implemented as "structural typing". [10:08] MikeBennett: @Matthew it would certainly make sense to define behaviour as one facet among many, under which things might be classified. In FIBO we've been exploring how to do metadata to identify separate classification facets, as there are so many different ways to classify the same thing. [10:08] PascalHitzler: GeoVoCamps mentioned by Werner: See e.g. http://vocamp.org/wiki/GeoVoCampSB2014 [10:08] PeterYim: == AldoGangemi presenting ... [10:09] GaryBergCross: Pascal +1 on mentioning the upcoming Vocabulary camp at Santa Barbara this March. People should come if interested in what they hear today. [10:12] GaryBergCross: Why ants on this 2nd slide??? [10:12] MatthewWest: Why not? [10:13] SimonSpero: @Gary: It's not an ant, it's a feature [10:41] AldoGangemi: @Gary ants were there to hint at social semantics, but @Simon I like them as features as well :) - [10:13] KarlHammar: For chairs: it seems the fire alarm at my university just went off. I need to figure out whether to vacate the premises or whether it is a false alarm. - [10:14] PascalHitzler: thanks for the notice Karl - hope it's nothing. - [10:14] KarlHammar: Trying to relocate to the library building, hoping it will not mess up my presentation here. Apologies. - [10:15] KrzysztofJanowicz: thanks for letting us know Karl, please keep us up to date. - [10:18] KarlHammar: Krzysztof: seems to have been a false alarm and I am now back. - [10:18] KrzysztofJanowicz: okay, thanks [10:15] DonnaFritzsche: I agree with - "What we need is an entity-centric frame-oriented data science to ensure relevance" [10:17] anonymous morphed into MarcelaS [10:17] PeterYim: @MarcelaS - kindly morph into your full name (in upper camel case) if you please - [10:17] AmandaVizedom: Double-checking: are we at slide 13 now? - [10:18] FrankLoebe: 14 I think - [10:19] KrzysztofJanowicz: 15 now - [10:20] KrzysztofJanowicz: @Peter: we are on slide 15 now - [10:20] FranLightsom: 21 [10:21] anonymous19 morphed into DonnaFritzsche - [10:26] SimonCox: Aldo's verbal numbering is not synced with numbers on slides in download stack - [10:26] JimSolderitsch: Slide 28 - [10:26] AliHashemi: His animations were flattened, there's an offset of 4 from what he says. - [10:33] KrzysztofJanowicz: slide 38 - [10:34] AliHashemi: i believe it is slide 42 or 43 in the pdf version [10:36] KrzysztofJanowicz: [even the vnc slide display may be off; run them on your on desktop by downloading from:] http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2014_02_06 [10:36] LianaKiff: I get an error when following the link to Fred. [10:36] LianaKiff: http://wit.istc.cnr.it/stlab-tools/fred/ [10:39] GaryBergCross: @Liana - I got to the Fred! page from the link. [10:40] AldoGangemi: @Liana on Firefox and Safari it works, I don't know on other browsers [10:41] GaryBergCross: The Fred link worked on Chrome. [10:42] LianaKiff: Tried Firefox. I'll try Chrome instead. [10:43] LianaKiff: Chrome worked. Thanks. [10:36] PeterYim: == KarlHammar presenting ... [10:43] JohnAYanosyJr2 morphed into JohnAYanosyJr - [10:43] PeterYim: ... now on slide#6 [10:46] PascalHitzler: I'm afraid I have to leave the session for teaching. Thanks for joining the call! [10:47] PeterYim: Bye, Pascal ... thank you for putting this together [10:49] AldoGangemi: (wave) to Pascal [10:49] SimonSpero: [cyclomatic complexity - for instances, not classes, right] [10:50] PeterYim: KarlHammar: ... noted typo on title of slide#11 (should be "Average Class In-Degree" instead of "Avgerage ...") [10:55] KrzysztofJanowicz: GCI (General Concept Inclusion): depends on whether you would like surface semantics or a deeper semantifications. we use GCI in our design patterns. [10:57] KrzysztofJanowicz: IMHO the problems are the tools, not the GCI [10:59] AliHashemi: I wonder if there is an inherent bias in some of the underlying performance metrics and benchmarking. While it is true that the naiive application of reasoners to large ontologies may yield significant slow-downs, query rewriting, reasoning federation and axiom-relevancy ranking + pruning can mitigate some subset of the performance problems. Are there metrics that take into account such factors on reasoning performance? - [11:00] KrzysztofJanowicz: Karl: 2 min left [11:00] AmandaVizedom: @KarlHammar -- related to @KrzysztofJanowicz's comment, it does seem that these recommendations need to be contextualized to language and tooling at least, yes? That is, the performance aspects are specifically consequences of representation pattern + language + implementation. Very different performance profiles might result from the same generally-described patterns otherwise implemented. [11:01] DonnaFritzsche: I agree with AmandaVizedom - good points [11:03] AmandaVizedom: re my 11:00 commment: In other words, the observations could be equally well expressed as constraints of the KR language: "If you us language X and reasoner(s) Y, for best reasoning performance you should observe these patterns..." Yes? [11:01] PeterYim: == Q & A and Open Discussion ... [11:01] AndreaWesterinen: Regarding documentation and several of the points made in this session, they overlap with the concepts discussed in Track A. I tried to summarize these points on our Track A community input page. http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014_Common_Reusable_Semantic_Content_CommunityInput [11:01] anonymous morphed into amin - [11:05] PeterYim: @amin, please morph into your RealName (in WikiWord format or upper camel case) ... this is the convention here and for proper attribution purposes [11:02] SimonSpero: Thanks to the speakers - have to leave for another call [11:02] anonymous morphed into UriShani [11:03] anonymous1 morphed into LamarHenderson [11:01] MatthewWest: [regarding the KarlHammar talk] I would want to challenge that the primary judgement of an ontology being "good" is that it performs well in a reasoning engine. I think there are many other things to consider, like that it does actually contain the intended model as a possible model, and perhaps even minimises the number of possible models. [11:03] MikeBennett: @Matthew I would take that a step further and ask whether the judgement of "good" as it applies to reasoning engine applications, also apply in other uses of formal semantics, e.g. in Big Data, conceptual ontologies (for data integration etc.). Is reasoning a pre-requisite (for instance, because you then know that the meanings are consistent and complete)? [11:03] MatthewWest: Don't over specify is a good, but not new point. Good to see that this is still the main principle though. [11:04] AliHashemi: [ref. verbal question from MichaelGruninger] Question from Slide 3 of chair: How to arrive at reusable patterns? How many patterns are there? Are there types of patterns? Are all patterns domain-independent? Can we mine patterns from data? [11:05] MichaelGruninger: @Krzysztof: Which Dagstuhl Seminar are you referring to? [11:06] KrzysztofJanowicz: @Michael: http://www.dagstuhl.de/en/program/calendar/semhp/?semnr=12221 [11:06] SimonSpero: @Michael: http://www.cs.helsinki.fi/u/langohr/graphmining/ [11:06] MatthewWest: I'll try to answer the 1st question. There are an unlimited number of patterns, because there are an unlimited number of atomic elements. Some patterns at least are domain dependent. Yes we can mine patterns from data, in fact this is one of the best ways to develop patterns. [11:06] ToddSchneider: Without trying to be facetious, what is a pattern? How can one be identified? [11:07] UriShani: @KarlHammar: Indeed, the guidelines you specify sound like good guidelines for writing ontologies. Is there a real distinction between ontologies and ODPs per this aspect? [11:09] KarlHammar: @UriShani: Good point. I suppose that pattern rewriting is easier than full ontology rewriting, but in principle, the design guidelines would seem to carry over also to ontologies. [11:03] GaryBergCross: @KarlHammer, Based on your indicators what is an example of a "good" pattern from the ones you looked at? [11:07] KarlHammar: @Gary: Haven't looked at the relative performance penalties of the various indicators, which would be necessary to calculate which existing patterns are "best" in terms of performance. [11:09] KrzysztofJanowicz: Image schema: +1 [11:09] TorstenHahmann: @KarlHammer: can we even evaluate patterns independently of the ontology they are used in? I can imagine that a pattern may perform well in the context of one particular ontology but not in another. [11:12] KarlHammar: @TorstenHahmann: Given how ODPs are are generally used (very simplified view: find/specialize/combine), it makes sense to consider the performance both of the ODP in itself, and the performance of the ODP when combined with other ODPs or existing ontologies. The latter issue is only briefly touched upon in this work, but I agree that it needs to be considered. [11:09] MatthewWest: 2nd Question. For true patterns, they will mostly be discovered, rather than invented. In the end, standards organizations will curate them. Good patterns are always useful, because they save effort and improve quality. The more a pattern is used the better it gets as bugs are eliminated. [11:14] KarlHammar: Agree with Matthews answer to 2nd question above. ChrisWelty did a keynote at Workshop on Ontology Patterns at ISWC 2010, touching upon exactly this. He called it "pattern archeology", i.e. the "digging up" of patterns from established systems/practices/models/etc. A process of discovery as opposed to design. Perhaps the keynote is available in WOP proceedings. [11:10] KrzysztofJanowicz: Kuhn's vision statement 'Modeling vs Encoding for the Semantic Web': http://www.semantic-web-journal.net/sites/default/files/swj35.pdf [11:12] MatthewWest: 3rd Question: when you abstract from ontology designs you are usually moving up the subtype/supertype hierarchy rather than moving out class-instance, so you should not normally need another language. Buy in comes from utility plus ease of availability and use. [11:13] GaryBergCross: Can our panelists talk a bit more on how "composability" is built/designed into a pattern? [11:14] MatthewWest: @Gary: What do you mean by "composability"? [11:16] GaryBergCross: @Mathew Composability was mentioned by Werner who can clarify from the Sortal example. [11:18] KarlHammar: @GaryBergCross: My answer above concerns composability in a more practical perspective, using OWL Content ODPs. As I interpreted Werner's example, this was a more general discussion, so my answer above is likely simplistic. [11:13] UriShani: @KarlHammar: where can I find an example of writing an ontology using ODPs? I use Protege. Is this the tool you will recommend? [11:21] AldoGangemi: @UriShani there is tool for ODP-based design, it is called XD (from eXtreme Design), it was implemented as an Eclipse plugin to the NeOn Toolkit tool [11:21] KarlHammar: @UriShani: Protege is fine. There are tools developed by STLAB, XD Tools, intended to support ODP usage. These tools are however intended for the NeOn Toolkit, which is not very well supported these days. I am considering porting these tools to Protege, but don't take that as a promise :) [11:22] AldoGangemi: we have no plans to port it to Protege, but anyone volunteering would be a great event! [11:22] KarlHammar: (actually, they have been halfway ported already by a master student, but may need a bit of polish) [11:22] AldoGangemi: @Karl that'd be great [11:13] AnatolyLevenchuk: At Track E Hackathon we already have at least one project that mention pattern search (raw data - mapping - ontology-linked data - pattern search - Linked Data web page presentation): http://ontolog.cim3.net/forum/ontology-summit/2014-02/msg00065.html [11:15] KarlHammar: @GaryBergCross: An interesting question which I'm actually aiming to adress in my future PhD work. I'm thinking Ontology Matching / Alignment methods / tools can be reused to aid in this. That would in turn require we give careful thought to labeling in ODP modules. [11:19] MikeBennett: (Summarizing @Aldo's verbal remarks) Complete repository of archetypical patterns... (primitives). Questions as to whether this is feasible. My take on this: maybe feasible in a simpler domain like business, more challenging if pursuing notion of archetypes for all human experience (per Leibniz etc.). I'm hearing confidence that the former at least can be done :) See also DOLCE. [11:20] AldoGangemi: @Mike good summary :) [11:21] MikeBennett: @Aldo Thanks - glad I captured it OK. This is something I am very motivated about. [11:19] GaryBergCross: I take this discussion of Gestalt primitives or image schema to be a relatively small group and that many other patterns are composites formed with them as parts. [11:22] GaryBergCross: @UriShani's [11:13] Question - in VoCamps we often start with a conceptual modeling tool like CMAP which can output a form of OWL. [11:23] AldoGangemi: @GaryBergCross @UriShani I highly recommend using things like CMap for early design with experts [11:23] KarlHammar: See http://sourceforge.net/p/extremedesign/code/HEAD/tree/trunk/xd-protege/ for the XD for Protege port. But again, this is early work in progress developed for a master thesis by one of Eva Blomqvists students, so it may still need some work to get running properly [11:24] AldoGangemi: it's close to what I called the "concept graph façade" of a KP [11:25] GaryBergCross: Karl's indicator factors may be used as a review of draft ODPs during a quality assessment phase. [11:25] UriShani: @all - thanks for the references and recommendations. [11:26] AldoGangemi: @Gary #Gestalt I agree it is a great working hypothesis, we need to prove it with large scale evidence (text and data grounding) [11:26] AnatolyLevenchuk: ISO 15926 community is very active in patterns related experiments. E.g. .15926 Editor supports pattern description and usage -- http://techinvestlab.ru/dot15926Editor [11:26] AldoGangemi: @Anatoly thanks for pointing, I'll look into it [11:27] KarlHammar: Thank you all for an interesting session. I'll definitely try to stay active in this interesting community! [11:27] AldoGangemi: Thanks Jano, Pascal, Peter, matthew [11:28] PeterYim: Please mark you calendars and reserve this time, every Thursday, for the OntologySummit2014 virtual panel session series. In particular ... next Track C session ill be on Thu 2014.03.20 (session-10) Session-05 will be up next Thursday - Thu 2014.02.13 (same time) - Track D: Tackling the Variety Problem in Big Data - I - see developing details at: http://ontolog.cim3.net/cgi-bin/wiki.pl?ConferenceCall_2014_02_13 [11:28] PeterYim: great session! [11:28] KrzysztofJanowicz: @KarlHammar: as example of GCI use http://geog.ucsb.edu/~jano/semantic_trajectories.pdf [11:30] KarlHammar: @Krzysztof: Interesting - printing it out now, and will check it out first thing in the morning (it is 20:28 here at the moment, I need to get out of the office ;)) [11:28] PeterYim: -- session ended: 11:25am PST -- ------